Romanian Journal of Morphology and Embryology 2006, 47(4):323 330 ORIGINAL PAPER Morphometric parameters and silver stain used in diagnosis of thyroid follicular diseases DELIA CIOBANU 1), IRINA-DRAGA CĂRUNTU 2), CARMEN VULPOI 3), NICULINA FLOREA 4), SIMONA-ELISA GIUŞCĂ 5) 1) Department of Pathology 2) Department of Histology 3) Department of Endocrinology Gr. T. Popa University of Medicine and Pharmacy, Iassy 4) Department of Pathology 5) I st Surgery Clinic Sf. Spiridon University Hospital, Iassy Abstract Fine needle aspiration biopsy (FNAB) of the thyroid is limited in distinguishing hyperplastic nodular goiter from true follicular neoplasm and in separating follicular adenoma from follicular carcinoma. The present study was done to evaluate if the morphologic and morphometric investigations and silver staining of nucleolar organizer regions (NORs), either alone or in association would help to differentiate the thyroid follicular diseases. 40 FNAB smears of thyroid follicular diseases, histopathologically diagnosed as nodular goiter, lymphocytic thyroiditis, follicular adenoma and follicular carcinoma, were analyzed using the standard cytological exam, quantitative analysis and NORs assessment. The qualitative evaluation, correlated with the numerical results obtained from the quantitative analysis, revealed that the cellular pattern, mean nuclear diameter and NORs area are valuable criteria in the diagnosis of the benign and malign follicular lesions, respectively. The results attained through morphometry increase the sensibility and the specificity of FNAB in the diagnosis of thyroid follicular carcinomas. Keywords: fine needle aspiration biopsy, thyroid follicular neoplasm, morphometry, NORs. Introduction During the last decade, the importance of fine needle aspiration biopsy (FNAB) has notably grown, coming to be considered the most accurate method in the management of thyroid nodular diseases. The percentage of patients guided to thyroidectomy has come down by 25%, while the cases of carcinoma that were surgically excised have increased by 15 30% [1 4]. The most difficult problem concerning the cytological diagnosis is making the difference between the follicular injuries (adenomas and follicular carcinomas). This is why these lesions are usually described as follicular neoplasms or thyroid follicular diseases, term introduced by Löwhagen T in 1979 [5], which includes hyperplastic nodules, follicular adenomas and follicular carcinomas. Beginning with the 80, many researchers have tried to increase the efficiency of thyroid FNAB s, by introducing new methods of evaluation (imunocytochemistry, morphometry or cytochemistry) [6 10]. The current study focus on the evaluation of thyroid follicular diseases (hyperplasias, adenomas, chronic lymphocytic thyroiditis and follicular carcinomas), through a panel of morphological, morphometrical and special stains (silver stain) for identification of the nucleolar organizer regions (NORs), in the FNAB. The originality of the study consists of using the same smears for the morphometrical and the cytochemical methods, the smears initially colored through the classical May-Grunwald Giemsa (MGG) method being discolored and argentically stained afterwards. Material and methods The study group was composed of 40 cases of thyroid follicular diseases diagnosed in the Department of Pathology of the Sf. Spiridon University Hospital Iassy during five years (1999 2003). The selection of the cases was based on the following criteria: all the cases were histopathologically confirmed as thyroid follicular diseases and were expelled from the cytological study the cases with oncocytic (oxyphilic) metaplasia (Hürthle cells) which showed obvious nuclear anomalies and inequalities, while excluding the errors of statistical analysis. We considered for each analyzed case: the morphological characters (the cellularity and the presence of the colloid), the dominant cellular arrangement ("honeycomb", syncytial, microfollicular, or isolated patterns), morphometric features (nuclear diameter and nuclear area [11]). The cellularity of the smears was classified in four categories: rich, moderate, poor and acellular smears,
324 Delia Ciobanu et al. depending on minimum quantity accepted in literature five or six cellular groups, with well preserved follicular cells, each group containing around 10 cells [1, 12, 13]. In the selected smears, we have enough cellular material for the morphological interpretation, for the morphometrical analysis and, further on, for the accomplishment of the cytochemical method. The 40 smears were standard colored with MGG stain. To evaluate the usefulness of the NORs stain, the smears were decolorized with a mixture of ethylic alcohol and 1% chlorhydric acid, fixed in ethylic alcohol 95% and rehydrated in alcohol 80%, for 15 minutes afterwards. There followed the stage of the washing in bidistilled dedurised water (10 minutes). The technique we applied was similar to the method of the silver stain on smears [14]. Additionally, we realized the coloring of the nuclei with methyl green for their better visualization. The microscopical images were acquired with a Nikon Eclipse 600 microscope, endowed with a Nikon DN 100 video camera, by using the Lucia Net software (running on the Department of Pathology, Sf. Spiridon University Hospital, Iassy). The images were made with 400x magnification, and the measurements were carried out on at least 15 20 cells per each microscopic field, with a total of 50 cells for each case. Concerning the silver stain, 50 cells were randomly chosen for each case and NORs were visualized with 100 magnification (with immersion). We counted the number of NORs per nucleus and computed the total NORs and nucleus Case Sex Age Cellularity Colloid Pattern areas, respectively. The marking of the nuclear contours was hand-made, and the measurement of the diameters and of the nuclear areas was made automatically, using the MICROIMAGE software (IMAGE ANALYSIS SOFWARE). The results obtained for each category of lesions were statistically processed with EXCEL and EPIINFO functions (t Student test, χ 2 test). Results Table 1 The studied cases of nodular goiter Synoptic presentation of the morphological and morphometrical characteristics of the studied groups The diagnosis of the 40 cases, based on the pathological exam, was nodular goiter 11 cases, adenoma 10 cases, lymphocytic thyroiditis nine cases and follicular carcinoma 10 cases. The analyzed cases of nodular goiter presented, from a qualitative point of view, rich or moderate cellularity of the smears (81.8%); the colloid was loose, abundant, palely basophilic, and represented the background of the smears in four cases (36.4%). The dominant cellular arrangement was the honeycomb pattern seven cases (63.63%), microfollicular three cases (27.27%) and had a syncytial pattern in one case (9.1%). We did not identify any isolated cellular arrangement. Table 1 presents, in addition, the qualitative aspects and the numerical results of the quantitative analysis. Measurement values for nuclei (pixels) diameter area ratio 1 f 74 1 2 honeycombs 107.9 ± 11.5 8515.9 ± 1598.5 78.8 2 f 41 2 3 honeycombs 90.1 ± 14.9 5370.6 ± 1701.8 59.7 3 f 44 2 1 microfollicular 61.8 ± 23.3 2751.8 ± 2011.3 44.4 4 m 36 1 1 microfollicular 88.5 ± 14.0 5087.2 ± 1488.3 57.2 5 f 53 2 2 microfollicular 88.7 ± 10.8 5415.4 ± 1251.2 60.8 6 f 56 3 3 honeycombs 100.0 ± 13.5 6788.1 ± 1834.5 67.9 7 f 45 3 3 honeycombs 118.2 ± 18.0 9400.9 ± 2971.8 79.7 8 m 41 2 3 honeycombs 96.6 ± 16.7 6040.1 ± 2307.1 62.3 9 f 64 2 2 honeycombs 104.0 ± 17.6 6898.1 ± 2032.2 66.3 10 f 58 3 2 syncytial 108.7 ± 22.5 8062.3 ± 4015.0 73.9 11 f 64 3 2 honeycombs 100.7 ± 16.9 6488.6 ± 2100.5 64.2 Total 95.4 ± 20.8 6253.2 ± 2710.6 65.8 The microscopic images, illustrating the morphometrical analysis, are presented in the Figures 1 and 2. By using the statistical methods of calculation in the cases of nodular goiter, the correlation coefficient had the value r = 0.92, which demonstrated the strong connection between the nuclear diameter and the nuclear area. We have also observed strong statistical correlation in all cases between the nuclear area and the NORs area, as well as between the number of NORs and their medium size (p < 0.001). In the analyzed cases of adenoma, the cellularity of the smears was rich or moderated in 90%, while the colloid was abundant or moderately present in seven cases. The most frequent cellular pattern was the microfollicular one (50% of the cases), followed by the one in "honeycomb" present in four cases (40%), and isolated in one case (10%). No case was identified in which the dominant arrangement of the follicular epithelial cells would be syncytial. Table 2 presents, in addition, the qualitative aspects and the numerical results of the quantitative analysis. The microscopic images, illustrating the morphometrical analysis, are presented in the Figures 3 and 4. In these cases, the value of the correlation coefficient (r = 0.94) demonstrated the strong positive correlation between the diameter and the area of the
Morphometric parameters and silver stain used in diagnosis of thyroid follicular diseases 325 nuclei. We obtained statistically significant correlation between the nuclear area and the NORs area, as well as Table 2 The studied cases of follicular adenoma Case Sex Age Cellularity Colloid Pattern between the number of NORs and their mean area (p < 0.001). Measurement values for nuclei (pixels) diameter area ratio 1 f 36 2 3 honeycombs 88.7 ± 10.5 5130.9 ± 1220.7 57.7 2 f 34 2 2 microfollicular 118.0 ± 19.6 8716.3 ± 2218.9 73.9 3 f 52 2 3 honeycombs 97.6 ± 11.8 5966.5 ± 1609.9 60.9 4 f 40 1 2 microfollicular 90.0 ± 9.2 5455.2 ± 810.5 60.6 5 m 64 3 2 microfollicular 100.5 ± 19.9 6704.3 ± 2174.9 66.4 6 f 45 3 2 honeycombs 106.2 ± 14.5 7758.8 ± 2224.4 73.2 7 f 63 3 1 isolated 83.4 ± 7.4 4929.5 ± 770.8 59.4 8 f 34 3 1 microfollicular 111.9 ± 9.4 8776.4 ± 1392.3 78.4 9 f 50 2 2 honeycombs 74.3 ± 8.1 3838.0 ± 840.0 51.9 10 f 35 2 1 microfollicular 106.2 ± 10.1 8001.8 ± 1996.6 75.5 Total 97.9 ± 18.6 6625.3 ± 2374.4 67.6 In the analyzed cases of lymphocytic thyroiditis, the smears situated themselves between rich (five cases 55.6%) or moderate (four cases 44.4%), while the colloid was abundant in 55.6% of the cases. The most frequent cellular arrangement was that of honeycomb Table 3 The studied cases of lymphocytic thyroiditis Case Sex Age Cellularity Colloid Pattern (five cases 55.5%), followed by the isolated pattern (two cases 22.2%) and by the microfollicular and syncytial cellular arrangement, one case each (11.1%). Table 3 presents, in addition, the qualitative aspects and the numerical results of the quantitative analysis. Measurement values for nuclei (pixels) diameter area ratio 1 f 37 2 1 honeycombs 99.7 ± 19.2 6669.8 ± 2664.8 66.7 2 f 36 2 2 microfollicular 67.7 ± 9.5 3020.5 ± 976.3 44.4 3 f 47 2 1 honeycombs 109.1 ± 20.5 8353.4 ± 3383.2 76.6 4 f 59 3 3 honeycombs 102.1 ± 13.9 7724.4 ± 1839.6 75.7 5 f 48 3 2 isolated 102.6 ± 14.8 7780.2 ± 1985.8 75.7 6 f 45 3 1 honeycombs 102.0 ± 10.0 7118.0 ± 1252.5 69.8 7 f 56 3 2 honeycombs 90.1 ± 8.3 5663.7 ± 1016.2 62.9 8 f 59 2 1 isolated 115.0 ± 12.7 9210.7 ± 1582.6 80.1 9 f 50 3 1 syncytial 100.2 ± 20.4 7211.3 ± 2969.1 72.1 Total 97.8 ± 17.9 6766.4 ± 2470.8 69.0 The microscopic images, illustrating the morphometrical analysis, are presented in the Figures 5 and 6. The value of the correlation coefficient, r = 0.94, demonstrated the strong correlation between the diameter and the area of the nuclei. We obtained statistically significant correlation between the nuclear area and the NORs area, as well as between the number of NORs and their mean area (p < 0.001). The 10 cases chosen to illustrate the malign thyroid Table 4 The studied cases of follicular carcinoma Case Sex Age Cellularity Colloid Pattern pathology presented either rich smears (six cases 60%) or moderate ones (four cases 40%). The colloid was missing in 30% of the cases, and the dominant cellular architecture was the syncytial one (five cases 50%), followed by the microfollicular pattern (three cases 30%). We have not identified the honeycomb arrangement in any case. Table 4 presents, in addition, the qualitative aspects and the numerical results of the quantitative analysis. Measurement values for nuclei (pixels) diameter area ratio 1 f 62 3 2 syncytial 112.9 ± 12.7 7785.8 ± 1425.4 68.9 2 f 21 2 1 microfollicular 103.3 ± 15.5 7553.0 ± 2105.2 73.3 3 f 70 2 0 syncytial 112.6 ± 10.4 8282.5 ± 1470.2 73.3 4 f 41 2 0 syncytial 110.7 ± 6.97 7900.1 ± 1212.1 71.2 5 f 22 3 2 syncytial 96.4 ± 10.3 6281.6 ± 1342.0 65.4 6 f 26 3 1 microfollicular 102.3 ± 8.17 7375.3 ± 1084.9 72.3 7 f 32 2 2 syncytial 104.2 ± 7.15 7883.6 ± 869.1 75.8 8 m 52 2 1 isolated 106.8 ± 10.5 8005.7 ± 1659.8 74.8 9 f 61 2 0 isolated 102.8 ± 11.8 7460.9 ± 1651.7 72.4 10 m 54 3 1 microfollicular 117.4 ± 12.0 9459.4 ± 1709.2 80.8 Total 108.9 ± 16.5 8087.0 ± 2507.8 74.2
326 Delia Ciobanu et al. The microscopic images, illustrating the morphometrical analysis, are presented in the Figures 7 and 8. In follicular carcinoma as well the correlation coefficient (r = 0.92) demonstrated the strong correlation between the diameter and the area of the nuclei. A positive correlation was observed (r = 0.34) between the area of the nuclei and the area of NORs, and an inverse correlation (r = -0.46) between the mean value of the NORs and their number. The calculation of the NORs area shows the converse proportionality between their number and the determined area. Discussions By using the t-student test, we revealed statistically significant differences between the tumoral pathology (follicular carcinoma) and other categories of thyroid nodular diseases. The statistical analysis showed the fact that the nuclear area and the nuclear diameter, in the cases of follicular carcinoma, are significantly larger than in the case of any other type of thyroid nodular diseases. The area of NORs and the nuclear area have showed the same statistical significance, while the number of NORs has showed a statistical significance only in the case of thyroid inflammatory diseases (Table 5). Evaluation of the personal results, in comparison with similar studies reported in literature Value of the quantitative analysis in the differential diagnosis of the thyroid follicular diseases The numerical results obtained through morphometry provided interdependency between the carcinoma and the rest of the thyroid nodular lesions. Even though the data of the specialized literature confirm a high level of accuracy for the FNAB in the thyroid diseases, there are still discussions concerning the incapacity of giving a final malignity diagnosis for the thyroid follicular lesions, only on the basis of the morphological characteristics [15]. Table 5 Statistical differences between carcinoma and other thyroid nodular diseases Measurement values Thyroid diseases Carcinoma (603) Nodular goiter (511) t-student; p Nuclei area (pixels) 35423 ± 11926 30515 ± 11482 6.18; p < 0.001 NORs area (pixels) 1533.1 ± 898.7 1065.0 ± 745.0 9.50; p < 0.001 NORs number 2.1 ± 1.0 2.2 ± 1.2 1.49; N S Mean NORs area (pixels) 864.8 ± 691.1 632.8 ± 534.6 6.31; p < 0.001 Carcinoma (603) Adenoma (676) Nuclei area (pixels) 35423 ± 11926 30760 ± 13523 6.55; p < 0.001 NORs area (pixels) 1533.1 ± 898.7 1048.6 ± 700.4 10.66 ; p < 0.001 NORs number 2.1 ± 1.0 2.0 ± 1.0 1.79; NS Mean NORs area (pixels) 864.8 ± 691.1 656.4 ± 602.9 5.71; p < 0.001 Carcinoma (603) Thyroiditis (545) Nuclei area (pixels) 35423 ± 11926 24746 ± 16446 12.48; p < 0.001 NORs area (pixels) 1533.1 ± 898.7 1116.7 ± 1374.3 6.01; p < 0.001 NORs number 2.1 ± 1.0 1.2 ± 0.5 19.56 ; p < 0.001 Mean NORs area (pixels) 864.8 ± 691.1 964.1 ± 849.3 2.16; p < 0.05 The multiple secondary changes founded in nodular goiter are well known. So, there are nodules composed only of large, cystic dilated follicles, with a flattened epithelium, while other nodules are densely and present, sometimes, solid areas as well. If the aspiration involved such areas, the differential diagnosis with a follicular adenoma or with a follicular carcinoma is difficult. The follicular adenomas, in their turn, are classified in four categories: macrofollicular, mediofollicular, microfollicular and trabecular, and the cytological material obtained through FNAB fluctuate according to their cytological type. The different histopathological aspects in the follicular carcinomas represent a nightmare for the cytologist. The carcinomas can be well differentiated, if formed out of "normal" follicles, or poorly differentiated, with frequent solid and trabecular areas. As in the follicular carcinomas the polarity is lost in the follicles, there appears a syncytial or microfollicular aspect. The great majority of studies on cellular distribution show the presence of the honeycomb pattern in the nodular goiter, in the thyroiditis and in the follicular adenoma, while it is missing in the smears that belong to the follicular carcinomas [16, 17]. In our study, the honeycomb pattern was present with a percentage of 63.63% in the nodular goiter, of 40% in the adenoma, of 55.6% in the thyroiditis and was absent in the cases of carcinoma, in which the syncytial (50%) and microfollicular (30%) patterns dominated. The data we obtained are similar to those reported in the literature [18]. The syncytial pattern represents densely cellular areas, made out of disorganized follicles. For this pattern, the information is contradictable: Kini SR et al. [16], concerning a studied series, do not report any case of adenoma with a syncytial pattern, while Ravinsky E and Safneck JR [17] find this pattern in all studied categories.
Morphometric parameters and silver stain used in diagnosis of thyroid follicular diseases 327 Figure 1 Nodular goiter (MGG, 400) Figure 2 Nodular goiter (silver stain, 1000) Figure 3 Follicular adenoma (MGG, 400) Figure 4 Follicular adenoma (silver stain, 1000)
328 DeliaCiobanu et al. Figure 5 Lymphocytic thyroiditis (MGG, 400) Figure 6 Lymphocytic thyroiditis (silver stain, 1000) Figure 7 Follicular carcinoma (MGG, 400) Figure 8 Follicular carcinoma (silver stain, 1000)
Morphometric parameters and silver stain used in diagnosis of thyroid follicular diseases 329 The results obtained in our study tend to be in accordance with Kini SR et al. [16]. In our opinion, the syncytial pattern can sometimes be an artifact due to an aggressive cellular manipulation, that can determine a cellular disorganization or characteristic superposition. The study of the diameter and of the nuclear area made by Boon ME et al. [19] and Kini SR et al. [16] has shown statistically significant differences between the malign tumoral diseases and the benign ones, our study having reached the same statistically significant results. Preceding studies, dedicated to NORs morphometry [20 22] have allowed the differentiation of thyroid benign and malign lesions, by the correlation of the nuclear area with the NORs area. Our data are partially in accordance with these references. In our study, the ratio between the NOR area and their mean number per nucleus has had an inverse correlation, meaning that the number of NORs has decreased while they were growing in size. For the NORs number, we have obtained statistically significant values (p < 0.001) only between the follicular carcinoma (2.1 ± 1.0) and the thyroiditis (1.2 ± 0.5). In our opinion, the absence of correlation in the NORs/nucleus ratio comes from the fact that, in the thyroid carcinomas, NORs have frequently appeared in-groups and bigger. Consecutively, the choice of the best microscopical image to measure the nuclear area has imposed a general preview, losing some of its accuracy. This is why we find it necessary to realize the quantitative analysis of NORs and nucleus, outside the count of NORs, so as to minimize the risks of falsely positive results. Conclusions The association of morphometry and cytochemistry in order to establish a significant preoperatory cytological diagnosis system for the follicular lesions suspected of malignity need the effort of the pathologist, but the present study demonstrates the effectiveness of this association. The results attained through morphometry increase the sensibility and the specificity of FNAB in the diagnosis of thyroid follicular carcinomas. The technique we propose allow the use of the same smears classically colored by MGG for the silver stain as well. The technique is elaborate, it is particularly recommended for clinically controversial cases, for patients with an unbalanced metabolism or whether the surgical intervention presents the risk for developing postoperatory complications. Although our conclusions cannot be generalized for the moment, as the evaluation of a larger number of cases is compulsive, the data are stimulating. References [1] GOELLNER J. R., GHARIB H., GRANT C. S., JOHNSON D. A., Fine needle aspiration cytology of the thyroid, 1980 to 1986, Acta Cytol, 1987, 31:587 590. [2] CARUSO D., MAZZAFERRI E. L., Fine needle aspiration in management of thyroid nodules, Endocrinologist, 1991, 1:194 199. [3] ROSAI J., CARCANGIU M. S., DELELLIS R. A., Tumors of thyroid gland. Atlas of tumor pathology, Armed Forces Institute Pathology, Washington, 1992, 31 275. [4] ***, AACE/AAES Medical Surgical Guidelines for Clinical Practice: Management of Thyroid Carcinoma, Endocr Pract, 2001, 7:203 217. [5] LÖWHAGEN T., GRANBERG P., LUNDELL G. et al., Aspiration biopsy cytology (ABC) in nodules of the thyroid gland suspected to be malignant, Surg Clin North Am, 1979, 59:3 7. [6] BACKDAHL M., WALLIN G., LÖWHAGEN T., WILLEMS J., Planimetric studies of fine needle aspirates from follicular adenoma and carcinoma of the thyroid, Acta Cytol, 1980, 24:145 148. [7] FERRER ROCA O., BALLESTER GUARDIA E., MARTIN RODRIGUEZ J., Morphometric, densitometric and flowcytometric criteria for the automated classification of the thyroid lesions, Analyt Quant Cytol Histol, 1990, 12:48 55. [8] MONTIRONI R., ALBERTI R., SISTI S. et al., Discrimination between follicular adenoma and follicular carcinoma of the thyroid. Preoperative validity of cytometry on aspiration smears, Appl Pathol, 1992, 7:367 374. [9] DEMICCO C., ZORO P., GARCIA S. et al., Thyroid peroxidase immunodetection as a tool to assist diagnosis of thyroid nodules of fine-needle aspiration biopsy, Eur J Endocrinol, 1994, 131:474 479. [10] POLYMOSI T., TOTH V., SAPI Z. et al., Diagnostic value of AgNOR method in thyroid cytopathology: correlation with morphometric measurements, Diagn Citopathol, 1996, 14:140 144. [11] CĂRUNTU I. D., Morfometrie computerizată microscopică în histologie şi histopatologie, Ed. Gr. T. Popa U.M.F. Iaşi, 2003, 99 134. [12] HAMBERGER J. I., HUSAIN M., Semiquantitative criteria for fine needle biopsy diagnosis: reduced false negative diagnosis, Diagn Cytopathol, 1988, 4:14 17. [13] NGUYEN G. K., GINSBERG J., CROCKFORD P. M., Fine needle aspiration biopsy cytology of the thyroid: its value and limitations in the diagnosis and management of solitary thyroid nodules, Pathol Ann, 1991, 26:63 91. [14] RUSCHOFF J., PRASSER C., PLOTON D. et al., Diagnostic value of AgNOR staining in follicular cell neoplasm of the thyroid, Am J Surg Pathol, 1993, 17:1281 1288. [15] ROSEN J. B., AZADIAN A., WALFISH P. G. et al., Ultrasound fine needle aspiration biopsy in the management of thyroid disease, Am J Surg, 1993, 166:346 349. [16] KINI S. R., MILLER J. M., HAMBERGER J. I., Cytopathology of follicular lesions of the thyroid, Diagn Cytopathol, 1985, 1:123 132. [17] RAVINSKY E., SAFNECK J. R., Fine needle aspirates of follicular lesions in the thyroid gland: The intermediate cell type smear, Acta Cytol, 1990, 34:813 820. [18] SUEN K. C., How does one separate cellular follicular lesions of the thyroid by fine needle aspiration biopsy?, Diagn Cytopathol, 1988, 4:78 81. [19] BOON M. E., LÖWHAGEN T., WILLEMS J. S., Planimetric studies on fine needle aspirates from follicular adenoma and follicular carcinoma of the thyroid, Acta Cytol, 1980, 27:145 148. [20] SOLYMOSI T., TOTH V., SAPI Z. et al., Diagnostic value of AgNOR method in thyroid cytopathology: correlation with morphometric measurements, Diagn Cytopathol, 1996, 14:140 144.
330 [21] SCHECHTMAN L., KOREN R., HOROWITZ A. et al., Diagnostic value of AgNOR staining in thyroid cytology, Anal Quant Cytol Histol, 1998, 20:187 197. Delia Ciobanu et al. [22] ARACHO PERULA E., ROLDAN VILLALOBOS R., BLANCO GARCIA F., BLANO RODRIGUEZ A., Objective differential classification of thyroid lesions by nuclear quantitative assessment, Histol Histopathol, 1997, 12:425 431. Corresponding author Delia Ciobanu, Assistant, MD, PhD, Department of Pathology, Faculty of Medicine, Gr. T. Popa University of Medicine and Pharmacy, 16 University Street, 700 115 Iassy, Romania; Phone +40232 215 350, Fax +40232 215 288, E-mail: deliaku@yahoo.com Received: January 5 th, 2007 Accepted: April 10 th, 2007